STHIfOLOaT or THE INDO-PACIFIC ISLANBS, 



i4| 



of thin fiimlly is tlie Dmvirirtn liildiil, Ahhmijrli tlie tificpg of the 

 oii.irimil Muii-Aiiam 2 are not very atrouf*, thfx nre ti+^t'iiled eiiouirh 

 to Ifiid uj* t« tlie c<*ndii,sioii that tlie dual had a fmm :*iniilar to thd 

 pp'Valent Hunnlaic Jj, or, in otlipr word:*, that tltis waa ilie (t>rin 

 cnrrent in Honlh Tihet at tiif era of rlie lir.sr srreiif. Ill tna lair mi;^-ra« 

 tion. Ahhouffh the lahiul prefix uas that n!"2, 4 and 5, mid the unifc 

 also had it, the .Mon-Anani G nnd lOO (//-Inraj t-ram) sliou' that a 

 unit, haviiifj- the ^rutfiirnl, iiKsslnar into the doiital aiuj r^iliilant, prefix was 

 early current. TUe Hltnriaii dsn, Tlinchu '"-^-'n i.< unotiier form uf the 

 unit with the same arch-aii; smmdarv fonn ofthit* prelis. 



The later slender forniSt Iwitli of the prefixsa and root?", connect mftny 

 flf the southerti ^jsteniN, although the ninveinents and ^jun-ial ImTowiiiira 

 di-^furb the flfrr^einent thiiH indm i'ti. The prnfiress ot" asjdrate, slender 

 and twitraetfd forms of the truttural prpfix k illu.-ntruted In the dual in 

 2 of Gm'o gi-u\, Mikir At-ni, 7 of Sintrjiho ^/-nit, ^"ams. i-ugit^ Kufsia 

 /o-niMn, t iaro imd Lluda .y-rd- Tins {rroup or niovemeiU centrei* in the 

 we.-iteru extremity of tUi! Goro-Sinjiftho hund. Another well marked 

 proiii", xvliich hri'iikn throiiffh thitthani or huA been broken tlneiij^h by it, 

 l^ distin^-nitJjed by similar eliaimes in the broad ftirm^ of the prefix, Jo 

 2 Ati^raini has /f//-iie, the fidjaeent Tfnf^-sii <i:roup «-nat, ((Jjaiunjj 

 /^/-iies). Itj 7 the Vuma dialet.ts liave f/ffi-^ sha-, w/-, the 'iVuir.sit ffr. 

 tha-^ Ui-, ChepunL; i*nd rfunwar cArt-. The on;_'inal ktt- is preserved m 

 the bro:ul Dophlii /w-nao;. In !) the Ynma frronn ha-* tJm-^ tti-, Chepuiiij 

 tu-y !4bindu chu-t Sin^'^pho tsi., tlaio sh-^ liudo ch-, 1ji (j the Yiima 'j:t,^ 

 Tenpta .irr, and l.ejichii have ta-^ Slrindu An!,', no-, Mikir th^, 



while Sinjrpho, Garo, Tukjia and CAw\m\y^ wUnn k- imA MoJi 



The unit 7 ((5, 1 or 1) is found in Cbineae nnd Lmi; in Thouhu, 

 Manyak mid Bhotiun ; in Clnmgrln, Le|ieliti, Mile!i!ioan;r, Myhu and 

 MikiV. The dual form (o, 2, or *2) is found in Tibet in llLopa and 

 Gyarnnjr (2), and iti ncarh' all the ShUithern dialeMls. It wiis the fonii 

 ot* the dinieet that pivf the nrevaleiit numerid ^eri»*.s to the Siuth, and 

 it corre^ijMJndH with the 2 of tliat dialirt, thus provintr that 7 nasi still qui- 

 nitry or dual at thf pmod of th*^ urfjil Tibetmn mifnitioif, or immediately 

 before it. No exaiiipb' of the* lull I'nnn \a nresiTved. qnintirj Kam- 



bojan p-mm pil (rj, 2) beloii^^a to the earlier l>raviro~Hiiu^kla.ic ?ijhtem 

 of the £3outh. 



Tlie priiieifMil inffrences hearinff on the historical relation of tlie Tllma- 

 Isic to tiie Ciniieae nuimMuly urea;* fullowrf. l&t. Theeurheat CliijifSi' dia- 

 IpjOt that pive nuuieralw toTiln-thad the primary fiutinral form of (lie unit 

 either alone or with lat^-r tornif*. It i^c*till mirn-nt in 10 of flurpa, in u 1 

 of liyaiuiifT, i" 7 Atimyak and in 1, G, 7, lU, 2U of s^ome southern 

 diidectf*. Thej'M i:< no direct or L'fiucln-iiive evidence that it vvud ev«r cur- 

 rent in 3 or o, piihrr in Tiltet or the r^onih ; although the olo!4e resemiilaiico 

 of ^onle of its torm- in 10 m the Chinese ftihilairt 10 — smd of the liitter, 

 atnl siiuilar current ci»- Himalayan sibilant forni:*^ to the Chinese 3 — • 

 snake it pi obabJe that the eurlie.st Chinese system of Tibet h d g-nt- 

 tufiil foriori thruu^'-hunt,. The 8 and 100 mn-it have hail similar f.-rin.-* of 

 tlie unit, and not the labial of the current Cbinese/rhe diuii in thi^avsteni 

 niujit have hiid tho lorni nj^ong, Jii(f:ui^ or ii^ifok n*j;ak ikc. 2d. The later 

 Cldnewe phase, vvhitdi was comtnimiitrited to Tiliet boiore the great mi- 

 grutiua to tiiu c^outh, Is very diiUiictlj' maiked by the foiuid of 3, 4 inid 8, 



